Configuration method and device, and computing system

Abstract

The invention provides a configuration method and a device, and a computing system, and belongs to the field of embedded computing. Wherein each computing unit of the computing system comprises at least two types of operators; the method comprises the steps of receiving a to-be-configured target algorithm; analyzing the calculation types contained in the target algorithm and the number corresponding to each calculation type; configuring a target operator for each target computing unit according to the computing types contained in the target algorithm and the number corresponding to each computing type; according to the target algorithm and the target operator corresponding to each target computing unit, determining a network structure formed by all the target computing units; and accordingto the target operator corresponding to each target computing unit and a network structure formed by all the target computing units, generating a configuration file of the computing system for realizing a target algorithm. Different types of algorithms are realized through the same computing system, and the parallel processing capability and the hardware resource utilization rate of the computingsystem are improved.

Description

technical field [0001] The invention belongs to the field of embedded computing, and relates to a configuration method, a device and a computing system. Background technique [0002] In recent years, with the increasing demand of embedded computing systems for target recognition of image / voice signals and intelligent task capabilities such as system autonomous decision-making, intelligent computing has become an important direction for the development of embedded computing systems. Deep neural network intelligent algorithms, which are widely used in high-performance server environments, have difficulties such as complex network structures, variable algorithm models, high computational complexity, and large parameter storage requirements. Existing computing systems provide a single type of algorithm, algorithm Poor adaptability and configuration flexibility. [0003] It can be seen that the existing computing systems have technical problems of poor algorithm adaptability and...

AI Technical Summary

Problems solved by technology

Deep neural network intelligent algorithms, which are widely used in high-performance server environments, have difficulties such as complex network structures, variable algorithm models, high computational complexity, and large parameter storage requirements. Existing computing systems provide a single type of algorithm, algorithm Poor adaptability and configuration flexibility

[0003] It can be seen that the existing computing systems have technical problems of poor algorithm adaptability and configuration flexibility

Images